IICRC and RIA are urging the industry to comment on the 2027 NAICS revision proposing code 624231, Emergency Restoration Services. Restoration currently has no dedicated federal classification, which is why headcount, payroll, and disaster-workforce data all vanish into janitorial and general-contractor buckets. Comment window closes August 12, 2026.
Why it matters
A dedicated NAICS code is the plumbing under every federal system that touches you, SBA loan eligibility, federal disaster contracting set-asides, insurance-industry data on your class of risk, BLS wage surveys, Census Bureau economic reports. Without it, restoration is invisible to policy that gets written about your work. This is a two-day window for the whole trade to buy itself a seat at every future federal table.
The read
A federal code will not get one supplement approved, and anyone promising otherwise is overselling. What it buys is being counted: wage data, loan eligibility, disaster contracting, and a class of risk insurers can actually see. The organizing matters more than the classification. File before the twelfth, and write something specific about your own payroll and disaster work instead of signing a template.
Chip Merlin's roadmap post names five current State Farm hail cases (Moore, Barlow: federal, N.D./W.D. Oklahoma; Hursh, Nida, Barnett, Oklahoma state courts) and lists the institutional discovery targets: the Wind/Hail Fire Model Enhancement Team (WHMET), the "Wind/Hail Playbook" and its training decks, Xactware triggers and management-approval routing, consultant work by Accenture, HAAG, and AccuWeather, and internal dashboards showing full-roof-replacement rates. The headline data point being litigated: full-roof approvals allegedly fell from 70.4% to 34.7% in one 2021 window.
Why it matters
The specific mechanisms carriers use to convert full-replacement roofs into partials (software triggers, weather-model consultants, management review gates) are being pried open by name in open court. If Merlin's clients pry the box open, every future denial letter is quotable back at the carrier. Roofing / storm shops get a template for what to preserve on their side of the paper trail: adjuster identity, software prompts observed, consultant names on the report.
The read
Discovery motions do not pay for shingles, and none of these cases will resolve soon. The value is the vocabulary. Naming the model team, the playbook, the software triggers and the consultants changes how a deposition or an appraisal conference goes. Start preserving your side now: adjuster names, screenshots of any software prompt you are shown, and every consultant named in a denial letter.
Allstate CEO Wilson confirmed the carrier is running "ALLIE," an internal LLM they say is aimed at "more accurate pricing" and "more accurate claims." Allstate says it uses "only internal LLMs" (i.e. not a third-party vendor endpoint) and acknowledges ALLIE is "not completely built or deployed." Everything else (what ALLIE writes, whether it can auto-reduce or auto-deny an estimate, which states it's live in, and whether a licensed adjuster reviews its output) is unstated in the disclosure.
Why it matters
Allstate isn't a top-3 property carrier the way State Farm is, but the disclosure pattern is the story. Carriers are announcing AI in claims without saying where a human reviews it. That's the exact fact pattern California's SB 1120 (denials made solely by automated tools) was written to catch, and Texas's June 2026 TDI bulletin on AI in claims handling flagged. If your next Allstate estimate comes back with a machine-flavored edit, ask on paper: was this reduction generated or approved by an unlicensed automated system, and when. That question changes the record.
The read
Half of these announcements are internal chatbots, so a press release proves nothing on its own. The disclosure gap is the story. No carrier is saying which reductions a licensed human approved, and that is exactly the question two state rules were written to force. Ask it in writing on every reduction letter: was this generated or approved by an automated system, and when.
Heritage's Q2 2026 numbers: 350,887 total policies (down 5% YoY), $62M net income (up 28.5% YoY), combined ratio 65% (an 8-point improvement). Florida count is down to 112,304 from a 2021 peak of 241,581, Heritage's whole US book has shrunk from about 625,000 policies in 2021 to about 351,000 now. Operating in 17 states, no explicit geographic exit, the strategy is selective underwriting.
Why it matters
Half of Heritage's Florida book is gone in five years and the remaining half is more profitable, that's a case study in a carrier making more money by writing less risk, exactly the pattern behind FAIR-plan blowouts and coverage-thinning across CAT-exposed states. If you work Heritage claims in FL, expect fewer of them, tighter scoping, and more customers whose neighbors quietly landed on Citizens.
The read
Florida is stabilizing is the line, and the profit numbers support it. Look at the shape. Half the Florida book is gone and the remaining half earns more, which is stabilization by writing less risk rather than by pricing the same risk sustainably. Your call volume depends on how many homes still carry coverage. Check which referral neighborhoods moved to the state insurer, and adjust where you market.
Reins is entering CORE Group's vendor network as the recommended provider for phantom stock and profit-sharing programs. The pitch: employee incentivization, succession planning, and key-talent retention without the owner giving up equity or operating control. Positioned as the anti-PE-rollup structural option, no fee tiers or ownership percentages disclosed in the announcement.
Why it matters
This is the first named, restoration-industry-endorsed alternative to selling to a rollup in the arc. Phantom stock lets you tie retention payouts to enterprise value the way an equity carve-out would, without carving out equity, that's directly relevant to any mid-size shop trying to retain a project manager, GM, or senior estimator who's been offered a promotion at a competitor. Whether it lives up to the pitch will depend on tax treatment and vesting triggers, neither of which the announcement addresses.
The read
Sell to a rollup or grind it out alone was always a false pair. Phantom stock has been standard in dental and home services for a decade, and it ties a key manager's payout to enterprise value without giving up a share of the company. It is compensation, not a growth plan. Ask about tax treatment and vesting triggers before you sign anything.
LEVLR launched Rapid Review on Aug 3. You upload your estimate PDF and the carrier's estimate PDF; the tool returns a prioritized breakdown of what to address first, the documentation supporting each line, and a prewritten adjuster response ready to edit and send. First review is free; pricing beyond that is undisclosed. No stated integrations with Xactimate or Symbility yet.
Why it matters
This is the estimator-side counterpart to what LLMs like ALLIE are doing on the carrier side, the same pattern-matching applied to your half of the paper fight. For a mid-size shop that can't afford a full-time supplement writer, a $0-to-try button on a supplement drafter is worth an afternoon.
The read
A free first review is customer acquisition, and a generic prewritten response is a starter draft rather than a finished supplement. It is still worth an afternoon. The tooling is running both directions now, and a shop that cannot afford a full-time supplement writer just got a cheap one. Run your next two disputed estimates through it, then compare the draft against what your best estimator writes.
Merlin's post uses Wilson v. 21st Century to argue that on wildfire smoke claims the insurer cannot use the policyholder's suggested sampling locations as artificial boundaries, the industrial hygienist's professional judgment governs where samples are pulled, per the joint IICRC/RIA/CIRI Technical Guide. Concealed spaces (attics, HVAC, crawlspaces) need to be tested even if not explicitly requested by the homeowner.
Why it matters
If you're doing IAQ testing on smoke claims, the record you build for why you tested where you tested is now a coverage-fight artifact. "The homeowner didn't ask about the attic" is not a defensible sampling plan; "the CIRI Technical Guide requires professional judgment on concealed spaces" is. Contractors should be quoting the CIRI Guide in their reports.
The read
The habit is to sample where the homeowner points and where the damage is obvious. Sampling location is the variable that decides the claim, and concealed spaces are where the argument gets won or lost. Write the reason into every report: attics, ducts and crawlspaces tested because professional judgment and the technical guide require it, not because someone asked. Cite the guide by name.
Ordinance or Law splits into Coverage A (value of undamaged portions that must be demolished), Coverage B (demolition and site clearing costs), and Coverage C (the cost delta of rebuilding to current code). The example: a Florida condo association was ~$7M short on rebuild after two hurricanes because they held one of the three, not all three.
Why it matters
On any code-triggered rebuild (older buildings, egress upgrades, sprinklers, energy code) you'll be quoting a scope the policy doesn't cover unless the policyholder bought all three. This is a pre-scope conversation: check declarations for A / B / C limits before you commit to a repair path, or you're writing an estimate the customer can't pay.
The read
Explaining coverage to a distraught homeowner is not your job, and staying in your lane is fair. Checking three limits on a declarations page is not explaining coverage, it is protecting an estimate you are about to write. A code-triggered rebuild with only one of the three leaves the customer short. Add the check to intake: undamaged portions, demolition, and increased cost of construction.
The 84% "open to AI" figure comes from the 2026 State of the Industry Report. The current-use side of the same survey: most companies use AI to draft emails, adjuster responses, or clean up letter language: not for operational integration (SOPs, estimating, retained institutional knowledge). Named vendors in the commentary: Claude (Anthropic), specifically Claude Projects for team accounts; Xactimate flagged for lacking open API connections; WorkWonders.ai (author's company) as the purpose-configured example.
Why it matters
Two-track story. Track one: the operator side is under-adopted, most shops using AI as autocomplete, not as an operating system. Track two: the carrier side (ALLIE, prior TPA AI, Verisk AI review) is ahead of the operator side, which is the exact asymmetry that costs supplement dollars. Closing the gap is not "buy ChatGPT": it's building a workspace on your SOPs, pricing rules, and voice so the AI writes like you.
The read
Using AI to clean up emails is not adoption, it is autocomplete, and most of the trade stopped there. The asymmetry is what costs you. Carriers are pointing the same technology at pricing and review while your side polishes letters. Take the tool you already pay for, load it with your own procedures, pricing rules and adjuster response library, then have it draft the next supplement.
Four consolidation datapoints in one week. PuroClean is targeting Greensboro and Durham for franchise expansion, the Carolinas triangle. RestoPros named Michael McAllister president, a leadership-team build that reads like a scaling / possibly-sale prep. GreenTec National announced a "boutique commercial" national launch, coast-to-coast. Extreme Supplies opened Tampa.
Why it matters
Four moves in a week isn't a coincidence; it's the shape of the summer. Franchise expansion, president hires, national launches, and market openings are all consolidation-adjacent signals in an industry that has run this pattern for three years running. If you operate in NC (Greensboro/Durham) or FL (Tampa), you have a new named competitor by the end of Q3.
The read
Four announcements in a week is not four healthy businesses, and the real scoreboard is who survives the year. The tell is velocity. Capital moved faster in the last sixty days than it did in the spring, which usually means buyers are positioning ahead of a rate cycle. If you are in the named markets, read any letter of intent harder than you would have in January.
Helixco announced Connect, positioned as an Emergency Response Profile + CRM built exclusively for restoration. Detail is thin in the announcement (no pricing, no integration list, no launch geography) but the category (restoration-specific ERP as opposed to a generic CRM adapted to the industry) is the signal.
Why it matters
Vertical software (as opposed to a Salesforce or HubSpot fit-out) has always been a slow arrival in restoration. If Connect is priced for mid-size shops (still an if), it's a real alternative to the "everyone in this trade uses [big-name platform]" dependency.
The read
Restoration-specific platforms launch every year and most do not survive the third one. Vertical software still beats a general tool bent into shape, when it is real. Judge it on three answers before the demo impresses anyone: named customers you can call, a working connection to the estimating system you already use, and total cost over three years.